Vehicle Emergency Collision Avoidance Under Limit Handling Conditions

It is of great significance to improve driving safety and reduce the injury of traffic accidents.The automatic emergency collision avoidance(ECA)system of vehicles can help the ego car to avoid or reduce traffic accidents,which is of great help to improve driving safety.The design principle of the traditional ECA system is usually conservative,which limits the vehicle mobility to a certain extent.In some extreme accident scenarios,the application of drifting technique can explore the potential of vehicle mobility and avoid an inevitable collision.The objective of this research is to apply drifting technique to the active emergency collision avoidance algorithm of passenger vehicles and to extend the collision avoidance capability in limit handling conditions.The research is divided into two parts,i.e.the trajectory planning and decision-making algorithm,and the trajectory tracking algorithm,respectively.Firstly,the trajectory planning algorithm of collision avoidance is proposed.For an example conflict scenario,the research summarizes three kinds of conventional ECA methods and proposes a drifting u-turn method based on drifting technique.A vehicle dynamic model is built for the ego vehicle,and a trajectory planning algorithm for the drifting u-turn method is designed based on optimal control framework.Combined with the motion prediction of obstacles in the scenario and the dynamic of ego vehicle model,the capability limits of the four ECA methods are analyzed,and the collision avoidance decision-making algorithm is concluded.The above trajectory planning algorithm is simulated and analyzed.Results show that the drifting u-turn collision avoidance method can still achieve collision avoidance in the limit scenario,while all the other three collision avoidance methods fail.Then,a tracking control algorithm is designed for the trajectory planned by the proposed drifting u-turn ECA method.Since during drifting,a combination of driving,braking and steering operations are used,and the vehicle is in a strongly nonlinear mode,common linear controllers are hard to track the trajectory.So,two nonlinear controllers are designed,i.e.one is based on nonlinear model predictive control,and the other is based on feedback linearization.Two controllers are validated by joint simulations.Finally,the tracking controller based on feedback linearization is further validated on a real vehicle.Experimental results show that the controller can track the target trajectory well.The proposed drifting u-turn method has the advantage of collision avoidance in extreme scenarios.The research presents a collision avoidance method by drifting u-turn,including trajectory planning and tracking algorithms.Through simulation and real vehicle validation,the proposed method can realize more extreme collision avoidance in limit cases,which expands the collision avoidance ability of passenger vehicles in dynamic scenarios with small available space.The research outcomes are expected to work as foundations of future developments of collision avoidance and high-maneuverability vehicle control systems.